On Thu, 2006-02-23 at 10:05 +0100, Jean-Marc Lasgouttes wrote: > >>>>> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes: > > Georg> This is a hack that will bite later. getChar is supposed to > Georg> return the contents of an inset if it can be represented as a > Georg> single character, not some arbitrary character. > > I think I agree. > > Georg> What you really want to know is "Can the cursor descend into > Georg> this inset?" We have InsetBase::editable() that exactly answers > Georg> this question, the only disadvantage is that this is not > Georg> implemented in mathed yet. > > Yes. > > Georg> The attached patch implements that and works for me as well. > Georg> What needs to be done if this should go in is to check for side > Georg> effects everywhere where editable() is used (fortunately only ~ > Georg> 3 places), and better testing, but I think that this approach > Georg> is prefereable to the getChar hack. > > Only MathNestInset needs this. All the others (including > MathHullInset) derive from that. Actually, you can just check for > asNestInset. > > The following works for me.
I think it is also logically more valid.
BTW get rid of the
if (it < end() && ...
condition! It is redundant.
- Martin
signature.asc
Description: This is a digitally signed message part
